Vanna AI

Vanna AI

Vanna is an open-source Python framework that turns natural language into accurate SQL. Point it at any database — Postgres, MySQL, Snowflake, BigQuery, SQLite — and any LLM provider (OpenAI, Anthropic, Gemini, Ollama), and users can query company data in plain English without writing a line of SQL. Vanna 2.0 adds Tool Memory: every successful query is stored in a vector database, so the agent learns your schema, business logic and query patterns over time instead of relying on manual training. Multi-turn conversation, per-user access control, chart generation and custom tools are built into the framework. Self-host the OSS core for free, or pay for the hosted cloud layer with observability, audit logs, file storage and SSO. Built for data teams and developers who want to embed BI-grade chat into their own products without vendor lock-in.

Analytics
Visit website
Added on
Nov 4, 2026
Open sourceMulti-databaseMulti-turnAccess controlTool MemoryAny LLM23.8k GitHub stars
Product information

Everything worth knowing about Vanna AI

The complete picture — from everyday features to the technical detail builders and IT folks dig for.

Core features

What it actually does

Natural Language to SQL
Let anyone on your team query databases in plain English — no SQL expertise required. The agent writes, executes, and returns results.
Works With Your Whole Stack
Multi-database out of the box: SQLite, PostgreSQL, MySQL, Snowflake, BigQuery, plus others via docs. Bring your own LLM provider — Anthropic, OpenAI, Gemini, or local Ollama models.
Multi-Turn Conversations
The agent holds context across follow-up questions instead of treating every prompt as a cold start. Ask, refine, drill down.
Per-User Access Control
Every user has specific permissions that flow through the entire system — control who sees which data at the row and query level, not just a login gate.
Tool Memory That Learns
Successful queries get saved and reused. Similar questions hit memory and adapt an existing query instead of re-deriving everything from scratch — no manual training needed.
Technical capabilities
Hosted Admin Layer
Optional cloud features on top of the OSS core: observability with real-time metrics and distributed tracing, audit logs, secure file storage, data retention, and a hosted vector DB.
Open Source Core, No Lock-In
Vanna 2.0 is an open source, user-aware agent framework (23.8k GitHub stars). Self-host with pip install vanna, then move to Cloud or Enterprise without rewriting code — same core framework across all tiers.
Must watch videos

Who uses it

Real use cases, no hype
Natural Language SQL
Let users ask questions in plain English and get executable SQL back.
Self-Service Analytics
Free non-SQL teammates from the data team bottleneck.
Embedded Data Agent
Ship a SQL chatbot inside your own SaaS product with per-tenant access control.
Auto-Generated Charts
Turn query results into Plotly visuals without extra wiring.
Warehouse Q&A Bot
Point it at Snowflake or BigQuery and let it learn your schema over time.
Governed Data Access
Enforce row- and role-level permissions while still allowing free-form querying.
What's great
  • 23.8k GitHub stars with active OSS momentum
  • Free open source core, no vendor lock-in
  • Returns SQL, dataframe, and Plotly chart in one call
  • Plugs into any LLM and any database
Technical strengths
  • Tool Memory reuses past successful queries instead of regenerating
  • RAG over DDL, docs, and question-SQL pairs in a pluggable vector store
  • Per-user permissions enforced through the execution layer, not just the prompt
Where it falls short
  • Accuracy collapses if your training data is sloppy
  • Complex multi-join queries still get hallucinated
  • Explorer plan caps you at 20 questions/day
  • Python-first; other language SDKs are thin
Technical limitations
  • Novel-question path burns tokens on schema investigation
  • No built-in warehouse cost guardrails — bad SQL hits your bill
  • Training data goes stale on schema drift; re-ingestion is manual
For technical folks

The deep-dive specs

Framework Type
User-aware AI agent framework connecting LLMs to databases
Memory Backend
Vector database (Tool Memory) for storing SQL queries and tool usage patterns
Supported LLM Providers
Anthropic, OpenAI, Gemini, Ollama
Supported Databases
SQLite, PostgreSQL, MySQL, Snowflake, BigQuery (others via docs)
Installation
pip install vanna (open source) or hosted at app.vanna.ai
Deployment Options
Self-hosted, Vanna Cloud, on-prem, VPC, private cloud

FAQ

Includes technical Q&A
Yes. All tiers run the same core framework, so you can self-host with OSS and move to managed cloud or enterprise deployment without rewriting code.

Traffic Insights

Monthly Visits

—

November 2026

Growth Rate

0%

vs last month

Dominance

0%

in Analytics

Top Country

—

Top Countries Breakdown

—

Growth Trend

Stable

Traffic has remained stable.

Honest pricing

No sneaky tiers, no “contact sales”

Here's who each plan is actually for — and where the hidden charges might hit.

Explorer
Most picked
$50/mo

Small teams of 2-3 people

  • ['20 questions/day'
  • 'Admin features included in either Vanna OSS or Vanna Cloud'
  • 'API available'
  • 'Same day email support']
Team
$500/mo

Small companies or large teams

  • ['300 questions/day'
  • 'Setup support'
  • 'Same day live support'
  • 'All Explorer features']
Enterprise
Custom

Large companies

  • ['Unlimited questions'
  • 'On-prem deployment support'
  • 'SAML SSO'
  • 'Admin API for user provisioning & management'
  • 'Advanced integrations'
  • 'Custom work']
View Full Plans & Pricing on Vanna AI

Official pricing, checkout, and plan upgrades are handled directly on vanna.ai.

Reviews

What the internet actually thinks

Data refreshed weekly. No paid placements.

Trustpilot
Not reviewed yet
G2
G2
Not reviewed yet
C
Capterra
Not reviewed yet
Product Hunt
Not reviewed yet
COMMUNITY COMMENTS

Leave a Comment

Help others make informed decisions. Your honest feedback shapes the community's understanding of this tool.

Tips for commenting

  • •Be specific about features you used
  • •Share real use cases and results
  • •Mention both pros and cons
  • •Keep it honest and constructive
to leave a comment

Your comment is published immediately.

Filter:
Sort:

No comments yet

Be the first to share your experience with this tool.

Alternatives

Not quite the right fit?

Here's what else is out there.